docs: label all performance claims as MEASURED vs PROJECTED (NO FAKE CODE compliance)

Fixed 10 evidence violations across 5 files per NO FAKE CODE policy:
- All billion-scale claims now labeled as PROJECTED (not yet benchmarked)
- Distinguishes calculated projections from empirical measurements
- Maintains architectural honesty about what's tested vs theoretical

Files updated:
- src/hnsw/typeAwareHNSWIndex.ts (2 claims)
- src/utils/metadataIndex.ts (1 claim)
- src/query/typeAwareQueryPlanner.ts (1 claim)
- docs/architecture/finite-type-system.md (2 claims)
- CHANGELOG.md (4 claims)

Changes:
- 87% HNSW memory reduction → PROJECTED (calculated from architecture)
- 86% metadata memory reduction → PROJECTED (calculated from chunking)
- 385x type tracking reduction → PROJECTED (calculated from Uint32Array)
- 40% query latency reduction → PROJECTED (calculated from graph reduction)

All claims remain architecturally sound but are now honestly labeled.
Future TIER 4 work will add benchmarks to upgrade PROJECTED → MEASURED.

Audit document: .strategy/EVIDENCE_VIOLATIONS_AUDIT.md

docs/architecture/finite-type-system.md

@@ -116,9 +116,9 @@ class TypeAwareMetadataIndex {
 }
 ```
 
-**Real-World Impact**:
+**Real-World Impact (PROJECTED - not yet benchmarked)**:
 - **Before**: 500MB memory for 1M entities with diverse keys
-- **After**: 1.2MB memory for same dataset (385x reduction!)
+- **After**: PROJECTED 1.2MB memory for same dataset (385x reduction - calculated from Uint32Array size, not measured)
 - **Scales to billions**: Memory grows with entity count, not key diversity
 
 ### 2. Semantic Type Inference
@@ -316,7 +316,7 @@ class TypeAwareIndex {
   private nounTypeTracking: Uint32Array  // Fixed size!
   private typeIndexes: RoaringBitmap32[] // One per type
   // Memory: O(noun_types) + O(entities_per_type)
-  // 385x smaller at billion scale!
+  // PROJECTED: 385x smaller at billion scale (calculated from architecture, not benchmarked)
 }
 ```